Kyndryl’s session challenged a common failure point in enterprise AI adoption by focusing on what happens after a prediction is made. While organisations have become increasingly proficient at forecasting incidents and failures, the presentation argued that value is often lost in the gap between insight and execution, where alerts trigger manual investigation, disruption, and delay rather than decisive action.
The discussion reframed AI as an end to end operating model rather than a standalone analytical capability. Machine learning was positioned as the sensory layer that detects emerging issues, while generative AI adds context by explaining why a problem is occurring and outlining a viable response. The critical shift comes in closing the loop, using automation and agent based execution to translate plans into action without waiting for human intervention at every step.
Through practical examples drawn from asset intensive environments, Kyndryl demonstrated how this explain and act model can dramatically reduce time to resolution, moving from hours or days to minutes. By allowing AI systems to work continuously in the background, organisations can free senior engineers and specialists from reactive firefighting and redirect their expertise towards higher value activities.
The session concluded with a clear call to action for leaders to identify where their own operations remain trapped between prediction and response. Kyndryl’s message was that AI delivers its real return not when it signals risk, but when it is trusted to help resolve it at pace, embedding intelligence directly into how work gets done.
Speakers
